Author: Vincent Diepeveen
Date: 07:10:34 09/04/02
Go up one level in this thread
On September 03, 2002 at 17:03:16, Jeremiah Penery wrote:
>On September 03, 2002 at 13:38:21, Matthew Hull wrote:
>
>>On September 03, 2002 at 13:07:54, Vincent Diepeveen wrote:
>>
>>>On September 03, 2002 at 13:05:30, Matthew Hull wrote:
>>>
>>>It gives the reason why i took a closer statistical look
>>>at his data. In order to find what his speed penalties were
>>>for processors. Then the fraud he committed was easy to find
>>>out.
>>>
>>
>>Look man, if all you got is that 13.00000 to prove your point, it's a pretty
>>thin thread to hang your argument on. It looks like you're really reaching.
>>
>>And don't worry that you can't duplicate the numbers with DIEP on some
>>supercomputer. If it's not a Cray, you won't be able to anyway.
>>
>>But if you ever get a quad XEON with interleaved memory, and still can't get to
>>1.7, then at least you're getting closer to a real argument.
>
>Perhaps you missed some of the threads from a while back (a year or so).
>Vincent has claimed to get >2.0 speedup on 2 processors before. I'm not sure
>why suddenly he changes this to 1.6 or whatever now. Seems to me he makes up
>whatever numbers he wants to 'prove' his points, because obviously whatever he
>says becomes a proof.
There are 2 'deep' diep versions
a) diep version 1999-2002 (juli)
b) diep optimized for NUMA
In both categories there is a substantial
difference in speedup between several versions:
a1) with dangerous extensions
a2) without dangerous extensions
a3) without dangerous extensions and with forward pruning
a3 always has a > 2.0 speedup for simplistic reasons that it is
doing a dubious search. It is this version which had also a 4.0 speedup
in 1999 and which searched 20 ply in endgame in 1999 at 4 processors.
After that i have again experimented with forward pruning bigtime over
the years, but the version a2 is very close to a 2.0 speedup. If you
look to node counts of crafty you will see it also needs less nodes
than a single cpu search.
Bob denies it, but never shows node counts. What Bob needs to do is both
proof 1.7 speedup and at the same time show the node counts for every ply
at every depth while claiming that 1.7 speedup.
Bob did do this for Cray Blitz obviously. However Cray Blitz used hardly
nullmove, so the chance that a branch gives a major cutoff in little nodes
is a lot less likely, whereas bob himself sees for crafty today already
a 2 ply increase using nullmove (i feel it is on average a LOT more
than 2 ply though for DIEP, because if i search with all extensions
turned on fullwidth i hardly get above 9 ply and only after billions of
nodes i get to 10 ply; only without dangerous extensions fullwidth a
ply or 10 is possible).
I will not comment too long on the 2 ply estimate for crafty, but let's
do a simple run here at bob's favourite position (so the one where he
has a good speedup), without knowing now what the speedup is and let it
run for 30 minutes exactly:
2r2rk1/1bqnbpp1/1p1ppn1p/pP6/N1P1P3/P2B1N1P/1B2QPP1/R2R2K1 b b7e4
moves: solution ; singlemove Bxe4!!
name : Bratko-Kopec.22
EPD Kit revision date: 1996.04.21
unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].
hash table memory = 384M bytes.
pawn hash table memory = 6M bytes.
pondering disabled.
noise level set to 1.
Crafty v18.15 (1 cpus)
White(1): ponder off
pondering disabled.
White(1): mt 2
max threads set to 2
White(1): setboard 2r2rk1/1bqnbpp1/1p1ppn1p/pP6/N1P1P3/P2B1N1P/1B2QPP1/R2R2K1 b
- -
Black(1): d
+---+---+---+---+---+---+---+---+
8 | | | *R| | | *R| *K| |
+---+---+---+---+---+---+---+---+
7 | | *B| *Q| *N| *B| *P| *P| |
+---+---+---+---+---+---+---+---+
6 | | *P| | *P| *P| *N| | *P|
+---+---+---+---+---+---+---+---+
5 | *P| P | | | | | | |
+---+---+---+---+---+---+---+---+
4 | N | | P | | P | | | |
+---+---+---+---+---+---+---+---+
3 | P | | | B | | N | | P |
+---+---+---+---+---+---+---+---+
2 | | B | | | Q | P | P | |
+---+---+---+---+---+---+---+---+
1 | R | | | R | | | K | |
+---+---+---+---+---+---+---+---+
a b c d e f g h
Black(1):
Black(1): analyze
Analyze Mode: type "exit" to terminate.
clearing hash tables
time surplus 0.00 time limit 30.00 (3:30)
nss depth time score variation (1)
starting thread 1
647 1 0.03 0.14 1. ... Rfd8
726 1-> 0.11 0.14 1. ... Rfd8
808 2 0.14 0.24 1. ... Rfd8 2. Bd4
254 2 0.19 0.20 1. ... Nc5 2. Nxc5 Qxc5
264 2 0.22 0.16 1. ... Ne5 2. Nxe5 dxe5
1393 2-> 0.28 0.16 1. ... Ne5 2. Nxe5 dxe5
1772 3 0.31 0.35 1. ... Ne5 2. Nxe5 dxe5 3. Qf3
667 3 0.34 0.15 1. ... Rfd8 2. Bd4 Nc5
5186 3-> 0.41 0.15 1. ... Rfd8 2. Bd4 Nc5
7001 4 0.44 0.14 1. ... Rfd8 2. Bd4 Nc5 3. Nxc5 dxc5
8945 4-> 0.50 0.14 1. ... Rfd8 2. Bd4 Nc5 3. Nxc5 dxc5
13438 5 0.53 0.25 1. ... Rfd8 2. Bd4 Nc5 3. Nxc5 dxc5
4. Be5
13130 5 0.64 0.19 1. ... Rfe8 2. Bd4 Red8 3. Rac1 Nc5
14561 5 0.67 0.18 1. ... Nh5 2. Qe3 Bf6 3. Bxf6 Nhxf6
4. Qf4
54388 5-> 0.78 0.18 1. ... Nh5 2. Qe3 Bf6 3. Bxf6 Nhxf6
4. Qf4
83921 6 0.89 0.16 1. ... Nh5 2. Qe3 Bf6 3. Bxf6 Nhxf6
4. Qf4 Rfd8
6 1.03 ++ 1. ... Bxe4!!
193346 6 1.09 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
193346 6-> 1.16 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
245644 7 1.28 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
265658 7-> 1.39 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
395031 8 1.55 -0.55 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc5
459422 8-> 1.67 -0.55 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc5
899258 9 2.09 -0.70 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bc3 a4
984089 9-> 2.24 -0.70 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bc3 a4
1541842 10 2.74 -0.44 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Bd4 Nfd7
1966413 10-> 3.17 -0.44 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Bd4 Nfd7
3933293 11 4.78 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bd4 Re2
7. b6 Rd8
4483252 11-> 5.38 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bd4 Re2
7. b6 Rd8
8338607 12 8.39 -0.48 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rcc8 6. Bd4 Nfd5
7. Rac1 Rfd8 8. Rxc8 Rxc8
16973942 12-> 15.38 -0.48 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rcc8 6. Bd4 Nfd5
7. Rac1 Rfd8 8. Rxc8 Rxc8
27883394 13 23.56 -0.63 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bd4 Rb8
7. Rac1 Rxb5 8. Rc7 e5 9. Rxd7 exd4
10. Nxd4
35691370 13-> 29.91 -0.63 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bd4 Rb8
7. Rac1 Rxb5 8. Rc7 e5 9. Rxd7 exd4
10. Nxd4
60295418 14 48.61 -0.40 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Bxf6 Nxf6 6. Rdc1 Rd8
7. Nc8 Nd5 8. b6 Re2 9. b7
136243127 14-> 1:48 -0.40 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Bxf6 Nxf6 6. Rdc1 Rd8
7. Nc8 Nd5 8. b6 Re2 9. b7
216294742 15 2:47 -0.51 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Bxf6 Nxf6 6. Rdc1 Rd8
7. Nc8 Bf8 8. b6 Re2 9. b7 Rb2
252416426 15-> 3:16 -0.51 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Bxf6 Nxf6 6. Rdc1 Rd8
7. Nc8 Bf8 8. b6 Re2 9. b7 Rb2
444862591 16 5:39 -0.46 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Rac1 Rfc8
7. Nd4 Rxc1 8. Bxc1 Nfd5 9. Nc6 a4
10. Bd2
520351692 16-> 6:38 -0.46 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Rac1 Rfc8
7. Nd4 Rxc1 8. Bxc1 Nfd5 9. Nc6 a4
10. Bd2
865441917 17 10:50 -0.50 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Rac1 Rfc8
7. Nd4 Nfd5 8. Rxc7 Rxc7 9. Nc6 a4
10. Rc1 f5
1063268606 17-> 13:24 -0.50 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Rac1 Rfc8
7. Nd4 Nfd5 8. Rxc7 Rxc7 9. Nc6 a4
10. Rc1 f5
1994721592 18 24:44 -0.57 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Rac1 Rfc8
7. Nd4 Nfd5 8. Rc6 Bf6 9. Rdc1 Bg5
10. Rxc7 Rxc7 11. Rc6
quit 18 31:00 26/33? 1. ... e5
Black(0): quit
execution complete.
D:\CHESS\crafty>
Now the same version without nullmove:
unable to open book file [./books.bin].
hash table memory = 384M bytes.
pawn hash table memory = 6M bytes.
pondering disabled.
noise level set to 1.
Crafty v18.15 (1 cpus)
White(1): mt 2
max threads set to 2
White(1): ponder off
pondering disabled.
White(1): setboard 2r2rk1/1bqnbpp1/1p1ppn1p/pP6/N1P1P3/P2B1N1P/1B2QPP1/R2R2K1 b
Black(1): - -
Illegal move: -
Black(1): d
+---+---+---+---+---+---+---+---+
8 | | | *R| | | *R| *K| |
+---+---+---+---+---+---+---+---+
7 | | *B| *Q| *N| *B| *P| *P| |
+---+---+---+---+---+---+---+---+
6 | | *P| | *P| *P| *N| | *P|
+---+---+---+---+---+---+---+---+
5 | *P| P | | | | | | |
+---+---+---+---+---+---+---+---+
4 | N | | P | | P | | | |
+---+---+---+---+---+---+---+---+
3 | P | | | B | | N | | P |
+---+---+---+---+---+---+---+---+
2 | | B | | | Q | P | P | |
+---+---+---+---+---+---+---+---+
1 | R | | | R | | | K | |
+---+---+---+---+---+---+---+---+
a b c d e f g h
Black(1): analyze
Analyze Mode: type "exit" to terminate.
clearing hash tables
time surplus 0.00 time limit 30.00 (3:30)
nss depth time score variation (1)
starting thread 1
647 1 0.05 0.14 1. ... Rfd8
726 1-> 0.10 0.14 1. ... Rfd8
807 2 0.14 0.24 1. ... Rfd8 2. Bd4
251 2 0.19 0.20 1. ... Nc5 2. Nxc5 Qxc5
261 2 0.22 0.16 1. ... Ne5 2. Nxe5 dxe5
1420 2-> 0.28 0.16 1. ... Ne5 2. Nxe5 dxe5
2146 3 0.32 0.35 1. ... Ne5 2. Nxe5 dxe5 3. Qf3
650 3 0.36 0.27 1. ... Nc5 2. Nxc5 Qxc5 3. Bd4
724 3 0.39 0.15 1. ... Rfd8 2. Bd4 Nc5
1759 3 0.46 0.14 1. ... Nh5 2. Bd4 Nf4
6808 3-> 0.49 0.14 1. ... Nh5 2. Bd4 Nf4
11023 4 0.53 0.16 1. ... Nh5 2. Qe3 Bf6 3. Nc3
7139 4 0.58 0.14 1. ... Rcd8 2. Bd4 Nc5 3. Nxc5 dxc5
33165 4-> 0.64 0.14 1. ... Rcd8 2. Bd4 Nc5 3. Nxc5 dxc5
52659 5 0.69 0.26 1. ... Rcd8 2. Qe3 Nh5 3. Bd4 Nc5
93240 5 0.77 0.22 1. ... Rfd8 2. Bd4 Nh5 3. Ng5 Ng3
115407 5 0.82 0.18 1. ... Nh5 2. Qe3 Bf6 3. Bxf6 Nhxf6
4. Qf4
163744 5-> 0.94 0.18 1. ... Nh5 2. Qe3 Bf6 3. Bxf6 Nhxf6
4. Qf4
229131 6 1.07 0.13 1. ... Nh5 2. Qe3 e5 3. g3 Bf6 4. Bc3
6 1.36 ++ 1. ... Bxe4!!
617961 6 1.42 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
617961 6-> 1.50 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
854974 7 1.75 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
1965002 7-> 2.61 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7
3204789 8 3.47 -0.55 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc5
8558951 8-> 7.49 -0.55 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc5
13537850 9 11.02 -0.70 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bc3 a4
39041323 9-> 30.74 -0.70 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bc3 a4
62109822 10 45.92 -0.44 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Bd4 Nfd7
179739712 10-> 2:14 -0.44 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rc7 6. Bd4 Nfd7
269928019 11 3:18 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bd4 Rd8
7. b6 Re2
793449429 11-> 9:37 -0.69 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bd4 Rd8
7. b6 Re2
1203681300 12 14:15 -0.44 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rcc8 6. Nd4 Rfe8
7. Nc6 a4
3625442403 12-> 44:17 -0.44 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Nxb6 5. Bd3 Rcc8 6. Nd4 Rfe8
7. Nc6 a4
5277808477 13 63:23 -0.63 1. ... Bxe4 2. Bxe4 Qxc4 3. Qxc4 Rxc4
4. Nxb6 Rxe4 5. Nxd7 Nxd7 6. Bd4 Rb8
7. Rac1 Rxb5 8. Rc7 e5 9. Rxd7 exd4
10. Nxd4
We see that a search of 15 seconds with nullmove takes 44 minutes
without. At tournament time controls the difference is already 3 ply.
After that difference is 5 ply, obviously caused by the better
branching factor.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.